Vendekagon Labs
@vendekagonlabs.bsky.social
Computational scientist, composer/producer, science fiction writer, and martial artist.
Pinned
My dark ambient project Hadamard Gate started from a simple conceit: just as a hadamard gate in quantum computing takes |0⟩ or |1⟩ qubits & puts them into superposition, I started w/classical instruments & orchestration, broke them up w/granularizers, & layered them w/time effects & distortion.
My dark ambient project Hadamard Gate started from a simple conceit: just as a hadamard gate in quantum computing takes |0⟩ or |1⟩ qubits & puts them into superposition, I started w/classical instruments & orchestration, broke them up w/granularizers, & layered them w/time effects & distortion.
June 20, 2025 at 2:13 PM
My dark ambient project Hadamard Gate started from a simple conceit: just as a hadamard gate in quantum computing takes |0⟩ or |1⟩ qubits & puts them into superposition, I started w/classical instruments & orchestration, broke them up w/granularizers, & layered them w/time effects & distortion.
Q: w/LLMs, how much is pattern recognition, how much is lossily compressed regurgitation? Try this experiment:
- ask ChatGPT (or another chatbot AI) for its best family-friendly dad jokes, put in a few prompts & attempts.
- search w/operator site:reddit.com for best dad jokes.
- compare the two.
- ask ChatGPT (or another chatbot AI) for its best family-friendly dad jokes, put in a few prompts & attempts.
- search w/operator site:reddit.com for best dad jokes.
- compare the two.
May 20, 2025 at 3:46 PM
Q: w/LLMs, how much is pattern recognition, how much is lossily compressed regurgitation? Try this experiment:
- ask ChatGPT (or another chatbot AI) for its best family-friendly dad jokes, put in a few prompts & attempts.
- search w/operator site:reddit.com for best dad jokes.
- compare the two.
- ask ChatGPT (or another chatbot AI) for its best family-friendly dad jokes, put in a few prompts & attempts.
- search w/operator site:reddit.com for best dad jokes.
- compare the two.
current reading. also the theme of my creative and scientific computation efforts in 2025.
December 28, 2024 at 5:43 PM
current reading. also the theme of my creative and scientific computation efforts in 2025.
the utility of any branded software design philosophy over its lifetime reduces down to the worst outcomes experienced when it gets cargo culted.
December 16, 2024 at 1:41 PM
the utility of any branded software design philosophy over its lifetime reduces down to the worst outcomes experienced when it gets cargo culted.
Watching most of tech go all-in on scammy, low effort "AI" has been one of the greatest disappointments of my life.
December 14, 2024 at 12:56 AM
Watching most of tech go all-in on scammy, low effort "AI" has been one of the greatest disappointments of my life.
all lifestyle philosophies are predictions. all predictions are wrong.
(with apologies to @stewartbrand.bsky.social )
(with apologies to @stewartbrand.bsky.social )
December 13, 2024 at 2:09 PM
all lifestyle philosophies are predictions. all predictions are wrong.
(with apologies to @stewartbrand.bsky.social )
(with apologies to @stewartbrand.bsky.social )
attention hacking is not ok. dark ux patterns are not ok. clickbait -> ad arbitrage is not ok. causing & concealing harm to children is not ok. manipulation is not ok. connecting scammers to victims is not ok.
we do not have to accept these things as necessary or inevitable aspects of technology.
we do not have to accept these things as necessary or inevitable aspects of technology.
December 12, 2024 at 2:33 PM
attention hacking is not ok. dark ux patterns are not ok. clickbait -> ad arbitrage is not ok. causing & concealing harm to children is not ok. manipulation is not ok. connecting scammers to victims is not ok.
we do not have to accept these things as necessary or inevitable aspects of technology.
we do not have to accept these things as necessary or inevitable aspects of technology.
> can you tell me a one sentence horror story?
npm install ai
npm install ai
December 11, 2024 at 4:50 PM
> can you tell me a one sentence horror story?
npm install ai
npm install ai
100% mine as well. tbh, the main thing I've learned in comparing my LLM experience vs. what others report, is how many devs are copying and pasting then modifying whatever glue code they find for working with a random lib or vendor api while never entering a flow state.
My experience in experimenting with LLM coding so far:
Me: Write code that frobnicates
LLM: Sure here is non-functioningXYZ.code!
Me: That doesn’t work because of XYZ
LLM: Sorry here’s non-functioningABC.code!
Me: Now that’s broken because of ABC
LLM: My bad here is non-functioningXYZ.code!
Me: Write code that frobnicates
LLM: Sure here is non-functioningXYZ.code!
Me: That doesn’t work because of XYZ
LLM: Sorry here’s non-functioningABC.code!
Me: Now that’s broken because of ABC
LLM: My bad here is non-functioningXYZ.code!
December 11, 2024 at 4:20 PM
100% mine as well. tbh, the main thing I've learned in comparing my LLM experience vs. what others report, is how many devs are copying and pasting then modifying whatever glue code they find for working with a random lib or vendor api while never entering a flow state.